Many real-life problems such as distributed meeting scheduling, mobile frequency allocation and resource allocation can be solved using multi-agent paradigms. Distributed constraint satisfaction problems (DisCSPs) is a framework for describing such problems in terms of related subproblems, called a complex local problem (CLP), which are dispersed over a number of locations, each with its own constraints on the values their variables can take. An agent knows the variables in its CLP plus the variables (and their current value) which are directly related to one of its own variables and the constraints relating them. It knows little about the rest of the problem. Thus, each CLP is solved by an agent which cooperates with other agents to solve the overall problem. Algorithms for solving DisCSPs can be classified as either systematic or local search with the former being complete and the latter incomplete. The algorithms generally assume that each agent has only one variable as they can solve DisCSP with CLPs using “virtual” agents. However, in large DisCSPs where it is appropriate to trade completeness off against timeliness, systematic search algorithms can be expensive when compared to local search algorithms which generally converge quicker to a solution (if a solution is found) when compared to systematic algorithms. A major drawback of local search algorithms is getting stuck at local optima. Significant researches have focused on heuristics which can be used in an attempt to either escape or avoid local optima. This thesis makes significant contributions to local search algorithms for DisCSPs. Firstly, we present a novel combination of heuristics in DynAPP (Dynamic Agent Prioritisation with Penalties), which is a distributed synchronous local search algorithm for solving DisCSPs having one variable per agent. DynAPP combines penalties on values and dynamic agent prioritisation heuristics to escape local optima. Secondly, we develop a divide and conquer approach that handles DisCSP with CLPs by exploiting the structure of the problem. The divide and conquer approach prioritises the finding of variable instantiations which satisfy the constraints between agents which are often more expensive to satisfy when compared to constraints within an agent. The approach also exploits concurrency and combines the following search strategies: (i) both systematic and local searches; (ii) both centralised and distributed searches; and (iii) a modified compilation strategy. We also present an algorithm that implements the divide and conquer approach in Multi-DCA (Divide and Conquer Algorithm for Agents with CLPs). DynAPP and Multi-DCA were evaluated on several benchmark problems and compared to the leading algorithms for DisCSPs and DisCSPs with CLPs respectively. The results show that at the region of difficult problems, combining search heuristics and exploiting problem structure in distributed constraint satisfaction achieve significant benefits (i.e. generally used less computational time and communication costs) over existing competing methods.

<p dir="ltr">We analyze the query complexity of multiple search problems.</p><p dir="ltr">Firstly, we provide lower bounds on the complexity of "Local Search". In local search we are given a graph G and oracle access to a function f mapping the vertices to numbers, and seek a local minimum of f; i.e. a vertex v such that f(v) <= f(u) for all neighbors u of v. We provide separate lower bounds in terms of several graph parameters, including congestion, expansion, separation number, mixing time of a random walk, and spectral gap. To aid in showing these bounds, we design and use an improved relational adversary method for classical algorithms, building on the prior work of Scott Aaronson. We also obtain some quantum bounds using the traditional strong weighted adversary method.</p><p dir="ltr">Secondly, we show a multiplicative duality gap for Yao's minimax lemma by studying unordered search. We then go on to give tighter than asymptotic bounds for unordered and ordered search in rounds. Inspired by a connection through sorting with rank queries, we also provide tight asymptotic bounds for proportional cake cutting in rounds.</p>

Spanning tree optimization problems are related to many practical applications. Several of these problems are NP-Hard, which limits the utility of exact methods and can require alternative approaches, like metaheuristics. A common issue for many metaheuristics is the data structure used to represent and manipulate the solutions. A data structure with efficient operations can expand the usefulness of a method by allowing larger instances to be solved in a reasonable amount of time. We propose the 2LETT data structure and uses it to represent spanning trees in two metaheuristics: mutation-based evolutionary algorithms and local search algorithms. The main operation of 2LETT is the exchange of one edge in the represented tree by another one, and it has O(&radic;n) time, where n is the number of vertices in the tree. We conducent qualitative and quantitative evaluations for 2LETT and other structures in the literature. For the main operation of edge exchange in evolutionary algorithms, the computational experiments show that 2LETT has the best performance for trees with more than 10,000 vertices. For local search algorithms, 2LETT is the best option to deal with large trees with large diameters. / Os problemas de otimização de árvores geradoras estão relacionados a muitas aplicações práticas. Vários desses problemas são NP-difícies, o que limita a utilidade de métodos exatos e pode exigir abordagens alternativas, como metaheurísticas. Um questão relevante para muitas metaheurísticas é a estrutura de dados usada para representar e manipular as soluções. Uma estrutura de dados com operações eficientes pode aumentar a utilidade de um método, permitindo que instâncias maiores sejam resolvidas em um período de tempo razoável. Propomos a estrutura de dados 2LETT e a usamos para representar árvores geradoras em duas metaheurísticas: algoritmos evolutivos baseados em mutações e algoritmos de busca local. A operação principal da 2LETT é a troca de uma aresta na árvore representada por outra aresta. Esta operação tem tempo de O(&radic;n), onde n é o número de vértices na árvore. Conduzimos avaliações qualitativas e quantitativas para 2LETT e outras estruturas na literatura. Para a principal operação de troca de arestas em algoritmos evolutivos, os experimentos computacionais mostram que a 2LETT possui o melhor desempenho para árvores com mais de 10.000 vértices. Para algoritmos de busca local, o 2LETT é a melhor opção para lidar com árvores grandes com grandes diâmetros.

This thesis presents new solution approaches for land leveling, using optimal earthmoving vehicle routing. It addresses the Shortest Route Cut and Fill Problem (SRCFP) developed by Henderson, Vaughan, Wakefield and Jacobson [2000]. The SRCFP is a discrete optimization search problem, proven to be NP-hard. The SRCFP describes the process of reshaping terrain through a series of cuts and fills. This process is commonly done when leveling land for building homes, parking lots, etc. The model used to represent this natural system is a variation of the Traveling Salesman Problem. The model is designed to limit the time needed to operate expensive, earthmoving vehicles. The model finds a vehicle route that minimizes the total time required to travel between cut and fill locations while leveling the site. An optimal route is a route requiring the least amount of travel time for an individual earthmoving vehicle. This research addresses the SRCFP by evaluating minimum function values across an unknown response surface. The result is a cost estimating strategy that provides construction planners a strategy for contouring terrain as cheaply as possible. Other applications of this research include rapid runway repair, and robotic vehicle routing. / Master of Science

Les enjeux du développement durable, le réchauffement climatique, la pollution dans les grandes villes, la congestion et les nuisances sonores, l'augmentation des prix de carburants, sont parmi des nombreux facteurs qui incitent les pays développés à l'innovation dans les transports publics. Dans ce contexte, l'introduction des systèmes de vélos en libre-service, au cours de ces dernières années, est une des solutions adoptées par de nombreuses grandes villes. Malgré leur succès fulgurant dans le monde entier, il existe peu d'études fondamentales sur ce type transport urbain. Pourtant, leur exploitation et leur management par des opérateurs soulèvent de nombreuses questions notamment d'ordre opérationnel. Dans ce contexte, cette thèse s'adresse aux problèmes d'ordonnancement et de rééquilibrage des stations de vélos en libre-service. Ce sont des problèmes cruciaux pour la qualité de service et la viabilité économique de tels systèmes. Le rééquilibrage consiste à redistribuer le nombre de vélos entre les différentes stations afin de satisfaire au mieux les demandes des usagers. Cette régulation se fait souvent par le biais de véhicules spécifiques qui font des tournées autour des différentes stations. Ainsi, deux problèmes d'optimisation difficiles se posent : la recherche de la meilleure tournée du véhicule de régulation (ordonnancement de la tournée) et la détermination des nombres de véhicules à utiliser (rééquilibrage des stations). Dans cette optique, les travaux de cette thèse constituent une contribution à la modélisation et à l'optimisation de performances des systèmes de vélos en libre-service en vue de leur rééquilibrage et leur ordonnancement. Plusieurs méthodes d'optimisation et ont été développées et testées. De telles méthodes incorporent différentes approches de simulation ou d'optimisation comme les réseaux de Petri, les algorithmes génétiques, les algorithmes gloutons, les algorithmes de recherche par voisinage, la méthode arborescente de branch-and-bound, l'élaboration des bornes supérieures et inférieures, etc. Différentes facettes du problème ont été étudiées : le cas statique, le cas dynamique, l'ordonnancement et le rééquilibrage avec un seul (ou multiple) véhicule(s). Afin de montrer la pertinence de nos approches, la thèse comporte également plusieurs applications réelles et expérimentations / In our days, developed countries have to face many public transport problems, including traffic congestion, air pollution, global oil prices and global warming. In this context, Public Bike sharing systems are one of the solutions that have been recently implemented in many big cities around the world. Despite their apparent success, the exploitation and management of such transportation systems imply crucial operational challenges that confronting the operators while few scientific works are available to support such complex dynamical systems. In this context, this thesis addresses the scheduling and balancing in public bicycle-sharing systems. These problems are the most crucial questions for their operational efficiency and economic viability. Bike sharing systems are balanced by distributing bicycles from one station to another. This procedure is generally ensured by using specific redistribution vehicles. Therefore, two hard optimization problems can be considered: finding a best tour for the redistribution vehicles (scheduling) and the determination of the numbers of bicycles to be assigned and of the vehicles to be used (balancing of the stations). In this context, this thesis constitutes a contribution to modelling and optimizing the bicycle sharing systems' performances in order to ensure a coherent scheduling and balancing strategies. Several optimization methods have been proposed and tested. Such methods incorporate different approaches of simulation or optimization like the Petri nets, the genetic algorithms, the greedy search algorithms, the local search algorithms, the arborescent branch-and-bound algorithms, the elaboration of upper and lower bounds, ... Different variants of the problem have been studied: the static mode, the dynamic mode, the scheduling and the balancing by using a single or multiple vehicle(s). In order to demonstrate the coherence and the suitability of our approaches, the thesis contains several real applications and experimentations